OCCASIONAL PAPERS THE MUSEUM TEXAS TECH UNIVERSITY NUMBER 77 19 MARCH 1982 RESOLVING A PHYLOGENY WITH MULTIPLE DATA SETS: A SYSTEMATIC STUDY OF PHYLLOSTOMOID BATS Michael L. Arnold, Rodney L. Honeycutt, Robert J. Baker, Vincent M. Sarich, and J. Knox Jones, Jr. All comparative biological data (in this study morphological, chromosomal, electrophoretic, and immunological) ultimately must be interpretable within a specific framework—the actual phylogeny of the set of organisms under study. An obvious corol¬ lary is that each data set should contribute toward the delineation of that phylogeny. The quality and magnitude of contribution, of course, will vary with the nature of the data set and its mode of evolution, because two conditions are necessary. First, a change must have occurred along the lineage under study and second, evidence of change must be observable. In this study, we were concerned with delineating phylogenetic relationships among the three families of New World bats cur¬ rently placed in the superfamily Phyllostomoidea—Noctilionidae (one genus, two species), the Mormoopidae (two genera, eight species), and the Phyllostomidae (some 49 genera and approxi¬ mately 139 species). In addition, we have examined intrafamilial associations in the Noctilionidae and the Mormoopidae. The genera Mormoops and Pteronotus classically have been placed either in the family Mormoopidae or (formerly) in the sub¬ family Chilonycterinae within the family Phyllostomidae. The systematic association of Mormoops with Pteronotus, suggesting a shared common ancestor for these genera following their separa¬ tion from all other taxa in the superfamily, has been recognized since the work of Dobson (1875). Noctilio, on the other hand, 2 OCCASIONAL PAPERS MUSEUM TEXAS TECH UNIVERSITY clearly has presented more of a problem to chiropteran systema- tists. Some workers have associated Noctilio with the Emballonu- roidea (Dobson, 1875; Trouessart, 1897; Miller, 1907; Simpson, 1945), whereas others (Winge, 1892; Smith, 1972; Patton and Baker, 1978) have indicated a mormoopid-phyllostomid associa¬ tion for these bats. Chromosomal change between the Mormoopidae and Noctilion- idae constitutes the least amount of divergence thus far docu¬ mented between two mammalian families (Patton and Baker, 1978). Chromosomal data have been interpreted as supporting a common ancestry for the two families (five synapomorphic ele¬ ments present) and for the two mormoopid genera (one synapo¬ morphic element present—Patton and Baker, 1978), with two rear¬ rangements distinguishing Mormoops from Pteronotus (Baker and Bickham, 1980). In order to understand better the evolution of the superfamily Phyllostomoidea, we have reviewed the resolving power and sys¬ tematic value of classical comparative anatomy, karyology (using G- and C-banding techniques), electrophoresis of various proteins, and albumin immunology. Furthermore, we have evaluated the extent to which each of the above analyses contributes to the delin¬ eation of an internally consistent phylogeny. Materials and Methods Electrophoretic Analyses All samples were assayed for 18 isozymes. The enzyme and pro¬ tein systems were malate dehydrogenase-1 and 2 (Mdh-1,2), phosphoglucomutase-1 and 2 (Pgm-1,2), lactate dehydrogenase-1 and 2 (Ldh-1,2), isocitrate dehydrogenase-1 and 2 (Idh-1,2), a- glycerophosphate dehydrogenase (a-Gpd), leucine aminopeptidase (Lap), peptidase (Pep), phosphoglucose isomerase-1 and 2 (Pgi- 1,2), glutamate oxalate transaminase-1 and 2 (Got-1,2), indo- phenol oxidase (Ipo), albumin (Alb), and hemoglobin (Hb). Tissue preparations, staining procedures, and enzyme designa¬ tions followed those of Selander et al. (1971). For examination of electrophoretic data, populations (see list of specimens examined for numbered localities) 1-7 ( Pteronotus par- nellii), 8-9 ( Pteronotus davyi), 12-13 ( Pteronotus personatus ), 15- 16 ( Mormoops me galop hylla), and 18-20 ( Noctilio leponnus) were grouped according to species. A cladistic analysis was performed on the data set by using the electromorphs as independent charac¬ ter states (Hennig, 1966). The two species of Noctilio represented ARNOLD ET AL. —PHYLLOSTOMOID BAT PHYTOGENY 3 an outgroup comparison for Pteronotus and Mormoops in the electrophoretic study. Any allele present in both Noctiho and spe¬ cies of Mormoops and Pteronotus was considered primitive for the two families. Further resolution of phylogenetic relationships, based on electrophoretic data, was made possible by using the iso¬ zyme systems for which no primitive allozyme was distinguishable from the outgroup comparison. This involved examination of the polarity of allozymic variation, with the assumption that the most common electromorph was primitive for the mormoopid- noctilionid clade and that all other variants were derived. The results from the use of this technique did not involve the rear¬ rangement of any relationships defined by the outgroup process; rather, they further refined relationships involving previously undefined lineages. Immunological Analyses The albumins of Mormoops megalophylla, Pteronotus parnel- lii, and Noctiho leponnus were purified according to the tech¬ niques of Cronin and Sarich (1975). Antisera to these albumins were prepared in rabbits (three to four Dutch Belted rabbits per albumin according to the schedule of Sarich, 1969), and each individual antiserum then was titered using microcomplement fixation (MC’F) and pooled in reciprocal proportion to its titer (Sarich and Wilson, 1966). Additionally, two antisera pools to the albumins of phyllostomid bat genera ( Macrotus, Vampyrum, Glossophaga, Carolha, and Desmodus) and pteropoid bat genera ( Syconyctehs , Pteropus, Dobsonia, Nyctimene, and Paranycti- mene) were used to provide estimates of the amounts of albumin change along the Mormoops, Pteronotus, and Noctiho lineages, as well as to test for relationships to the phyllostomids. All anti¬ gens used for cross-reactions with the different antisera were extracted from samples of whole serum or tissue diluents. Immunological cross-reactions (antigen-antibody reactions) for all comparisons were measured by the quantitative precipitin technique employed by Sarich and Wilson (1966) and Prager and Wilson (1971). The degree of cross-reaction was expressed quan¬ titatively as albumin immunological distance units (AID), with one unit being approximately equivalent to one amino acid sub¬ stitution (Prager and Wilson, 1971; Maxson and Wilson, 1974), Specimens Examined Specimens used in this study were collected at the following localities: Pterono¬ tus parnellii. — 1) 1 km. N Merida, Yucatan, Mexico, 2 females; 2) Guatopo 4 OCCASIONAL PAPERS MUSEUM TEXAS TECH UNIVERSITY National Park, Santa Crucita Campground, Venezuela, 2 females, 1 male; 3) 1 mi. N El Dorado, Sinaloa, Mexico, 3 males; 4) 2 mi. NE Rosario, Sinaloa, Mexico, 1 male; 5) 0.4 mi. E Hwy. 15 on road to Acaponeta, Nayarit, Mexico, 1 male; 6) 24.1 mi. N Rio La Union on Hwy. 200, Guerrero, Mexico, 1 male; 7) 0.2 mi. E Water- mount, Jamaica, 16 females, 5 males; Pteronotus davyi. —8) 15 km. N Altagracia de Orituco, Guarico, Venezuela, 1 female; 9) Tanetane, St. John Parish, Dominica, 7 females, 10 males; Pteronotus macleayii. — 10) St. Clair Cave, St. Catherine Parish, Jamaica, 15 females, 5 males; Pteronotus quadridens. —11) St. Clair Cave, St. Cath¬ erine Parish, Jamaica, 11 females, 10 males; Pteronotus personatus. —12) Tehuan¬ tepec, Oaxaca, Mexico, 5 males; 13) El Fuerte, Sinaloa, Mexico, 1 male; Mormoops blaimnllii. —14) St. Clair Cave, St. Catherine Parish, Jamaica, 20 males; Mormoops me'galophylia. —15) 24.1 mi. N Rio La Union on Hwy. 200, Guerrero, Mexico, 1 female; 16) 8.2 mi. S Pina Blanca on Hwy. 120, Queretaro, Mexico, 2 males; Noeti- ho albwentris. —17) 15 km. N Altagracia de Orituco, Guarico, Venezuela, 1 female, 1 male; Noctilio leporinus. —18) 0.2 mi. E Watermount, St. Catherine Parish, Jamaica, 5 females, 3 males; 19) mouth of Belham River, St. Anthony, Montserrat, 3 females, 4 males; 20) 1 mi. above mouth of Layou River, St. Joseph Parish, Dominica, 1 female. Results and Discussion Electrophoretic Analyses Only one isozyme (Lap) was monomorphic for all populations (Table 1). Nine of the remaining 17 isozyme systems yielded information concerning primitive and derived conditions using the outgroup criteria. The cladogram represented by Fig. 1 was produced by first using the data from these systems and then further resolving relationships based on the remaining allelic data as discussed in the section on methods. Allozymes present in the internodes (for example, Got-1 156 ) are synapomorphic for the spe¬ cies located above tljte internode, whereas allozymic characters located on branches ending in a single species (for example, Idh- l 100 ) are considered aufapomorphies for that species. Several phylogenetic relationships are suggested from the cladis- tical analysis of electrophoretic data (Fig. 1). First, a clade com¬ posed of the five Pteronotus species is defined by four synapo¬ morphic electromorphs* Within that assemblage, P. parnellii is separated from the ot|ier four species by the Idh-1 133 and Got-1 156 allozymes. The two species of Mormoops and the two of Noctilio are uni¬ ted by three and seven shared characters, respectively, with all of these electromorphs belonging to loci for which the outgroup criteria failed to discrijninate primitive and derived conditions. ARNOLD ET AL.—PHYLLOSTOMOID BAI PHYLOGENY 5 Table ]. AUozyme data for the 20 populations of mormoopid and noctihontd bats (see text for identification of numbered populations). The mobility of the most frequent allozyme in population 2 was arbitrarily designated 100 with all other allozymes being designated relative to this electromorph. A negative (—) sign indi¬ cates a cathodal mobility of the allozymes, whereas the lack of any sign indicates anodal mobility. Population Isu/yrnt- 1 2-7 8-9 10 11 12 13 11 15-16 17 18-20 Mdli-1 100 100 100 100 100 100 100 31 11 138 138 Mdh-2 -100 -100 -107 -100 -100 -100 -100 -100 -100 - 89 - 89 P K m-l 100 100(0.98) 30(0.02) 100 100 100 133 153 15# 153 153 153 Pgm-2 -100 -100 -100 -129 -100 -100 -100 -100 -100 -100 -100 I.dh-I 100 100 100 100 100 100 100 126 82 117 117 I.dh-2 -100 -100 -100 -100 - 89 -100 -100 100 -100 -100 -100 Idh-1 100 100 133 133(0.12) 171(0.58) 133 133 133(0.50) 95(0.50? 111 161 1 11 111 Idli-2 -100 -100 - 62 - 38 -108 -100 -100 - 88 - 38 - 38 - 38 a-Gpd 100 100 ■18 18 18 100 100 66 66 115 117 Pep 100 100 100 100 81 78 78 130 122 182 182 Pgi-1 83 100 110 110 157 110 110 187 187 180 180 Pgi-2 100 100 100 120 100 140 110 210 210 80 80 (. 01-1 100 100 156 156 156 156 156 100 72 100 200 Goi-2 -100 -100 -100 - 73 -100 -100 -100 -100 - 68 -100 — 55 Ipo -100 -100 -100 -100 -100 -100 -100 -100 -100 - 31 31 Alb 100 100 96 100 89 100 100 117 112 112 112 Hb -100 -100 -100 -100 -100 -100 -100 -100 - 67 -100 -100 Lap 100 100 100 100 100 100 100 100 100 100 100 Morphological Analyses Smith (1972) analyzed the morphological relationships between the families Mormoopidae, Noctilionidae, and Phyllostomidae as well as the inter- and intrageneric relationships of mormoopid taxa. Although Smith’s study was basically phenetic in its approach, he did evaluate a large array of qualitative characteris¬ tics, not only for the Phyllostomoidea but also for other New World Chiroptera. This set of data is unique for bats in that it provides an adequate base for cladistic analysis. Fig. 2 represents our interpretation of these qualitative characters, a description of which appears in the legend. In deriving the cladogram, we assumed that emballonuroids represent a valid outgroup. The resolved phylogentic tree (Fig. 2) presents several salient features. The families Phyllostomidae, Mormoopidae, and Noctil¬ ionidae share several synapomorphies and represent a unified clade, as suggested by Smith (1972). However, a closer association of the Noctilionidae to the Mormoopidae than to the Phyllostom¬ idae is not demonstrable. Although the Mormoopidae and Noc¬ tilionidae share a single derived character, a synapomorphic ele- 6 OCCASIONAL PAPERS MUSEUM TEXAS TECH UNIVERSITY o c o L - qj C OJ 13 3 O' Idh-l(IOO) I I I 1(100)* Idhrlf^)* ldh-2(- ldh-2(-108) : Ldh-2f-89)* O C O V qj f IdMUIJ* p g i-2nioi -c CL _o (T3 OJ £ Q. O o E Q- I Alb(96]* ldh-2(-62l* O 5 I Col-2(-68l* PepftMS Lcfh^noot* ldh-2i-Mi* Goi-1 (1561* ldh-1(1J3l* AlbllOO]* Pgm-inoar ldh-2(-100|* [t-Gpd(IOO) '>-Gpd(66i Pgi-mS?) Pgi-2(240l o Z I Pgi-1(160i Pgi-2l80) 1(30] * Pep|7B) IVp W' p g rn-2J-129)"' Md h- 21 -IIVi lllrl IMP MhG17;-* PVp082l 1 Alb?89)« Gnf-2(-73)* | ltlh-1(821 Mdh-1[341 Mdh-1[13B| Pgi-1 [157f 1 1 Ldh-111 261 lptil-31] L Hb[-67l* Popi 1301 Ldh-111471 Mdh-1 Wi Mdh-2(-89l Eig. 1.—Cladogram based on allozymic variants. Allozymes denoted by asterisk were resolved using the outgroup method. See text for identification of abbrevia¬ tions. ment also is present between the families Phyllostomidae and Mormoopidae. One synapomorphy places the genera Mormoops and Pteronotus together; another (shape of tragus) separates P. parnelhi from its congeners, with the other four species of Ptero¬ notus forming an unresolved tetrotomy. One and three synapo- morphic characters, respectively, unites the species of Noctilio and Mormoops . Karyotypic Analyses Chromosomal data as presented by Patton and Baker (1978) and Baker and Bickham (1980) add additional resolution to phylogenetic relationships within the Phyllostomoidea (Fig. 3). The primitive karyotype for the superfamily, as proposed by Patton and Baker (1978), consists of a 2n=46 and FN=60, essentially the karyotype of the phyllostomid species Macrotus waterhousii. Based on the assumption that the 2n=46, FN=60 karyotype is primitive for the Phyllostomoidea, the families Noctilionidae and Mormoopidae are linked by five synapomorphies (Robertsonian fusions). Patton and Baker (1978), using the rule of parsimony, indicated that the ARNOLD ET AL.—PHYLLOSTOMOID BAT PHYTOGENY 7 Eig. 2.—Cladogram based on morphological characters examined by Smith (1972): 1, Premaxillary bones complete and fused to maxillary and palatine bones; 2, Os penis absent; 3, Marked sella turcica; 4, Foramen ovale similarly situated; 5, Wartlike bumps and ridges on lower lip; 6, Lanceolate tragus; 7, Capitulum and trochlea lie in an intermediate position; 8, Supraglenoid fossa well developed; 9, Trochanter on proximal end of femur not distinct; 10, Tragus slender, pinnately- lobed; 11, Medial process moderate in length; 12, Capitulum and radius form ‘Tongue and groove”; central part of capitulum larger; 13, Tragus with secondary fold greatly increased in size and forming major part of structure; 14, Medial pro¬ cess shortened; 15, Well-developed tragus fold; 16, Well-developed secondary tragus fold. 8 OCCASIONAL PAPERS MUSEUM TEXAS TECH UNIVERSITY F of 21/14,17/9,13/8,22/3,18/12 Proposed Primitive For Noctilionoidea Fig. 3.—Cladogram based on chromosomal data from studies by Patron and Baker (1978), and Baker and Bickham (1980): F=fusion, H+=heierothromatir addition, I=inversion, PA—paracentric inversion. fusion events probably were synapomorphies inasmuch as several more events, including fissions preceded by several fusions, would have to be invoked to explain a noctilionid-mormoopid-like karyotype as primitive. A cladistic analysis to determine a primi¬ tive karyotype for the Phyllostomoidea cannot be performed at this time because homologous elements have not been identified with appropriate outgroups. Mormoopid and noctilionid species differ by four rearrangements (Patton and Baker, 1978)—a fission event in the mormoopids (synapomorphic for Mormoops and Pteronotus) and an inversion, a fusion event, and a heterochromatic addition in the noctilionids (synapomorphic for the two Noctilio species). Additionally, Mormoops blainvillii differs from the spe¬ cies of Pteronotus by a paracentric inversion and a heterochro¬ matic addition (Baker and Bickham, 1980). Immunological Analyses A salient feature of almost all phyllostomoid albumins is their immunological distinctiveness relative to those of other bats. The assumption of microchiropteran monophylly or the use of phy¬ logenetic analyses involving nonbat reference species (Sarich, unpublished data) leads to the conclusion that an event producing ARNOLD ET AL.—PHYLLOSTOMOID BAT PHYLOGENY 9 Table 2, —Albumin immunological distance values for mormoopid and not'd lion id species. The Phyllostomidae sample is a mixed outgroup consisting of Macrolus, Vampyrum, Glossophaga, Carollia, and Desmodus. The Pteropodidae sample con¬ sists of Syconycteris, Pteropus, Dobsonia, Nyctirnene, and Paranyctimene. Tamm Sl.rn. p.p. S.l. Ph Mormoops megalophylla 0 95 146 87 Pteronotus parnellii 0 150 80 Noctilio leporinus 0 135 Phyllostomidae 0 Pteropodidae 140 178 180 172 the antigenic equivalent of 30 to 40 units of albumin immunolog¬ ical distance must have occurred early in phyllostomoid history (Table 2, Fig. 4). Thus, this event has provided an effective phe- netic marker that positions a wide variety of New World bats into a single clade. It also would appear that such an event has caused a relative rate destabilization for subsequent albumin evolution in the group (at least as it is assessed immunologically). There is appreciably more variation in observed amounts of change along different lineages than has been demonstrated in any other verte¬ brate group (Honeycutt and Sarich, unpublished data). It thus is fair to point out that the “molecular clock" concept never would have been formulated on the basis of a study of phyllostomoid albumin evolution. This does not, however, affect our cladistic analysis, as the data for phyllostomoid bats apportion into addi¬ tive phylogenies at least as well as do those for any other group for which similar information is available. For example, the F value (Prager and Wilson, 1978) for the input-output comparisons involving 16 phyllostomid albumins and antisera to them is less than 5 per cent (Honeycutt and Sarich, unpublished data). As systematists, we find it disturbing that the cladrstjcal analy¬ sis of data from albumins of Mormoops are not consistent with the results of similar analyses of morphological and chromosomal data. Mormoops, Pteronotus, and various phyllostomid albumins are more or less equidistant from one another. Indeed, if any¬ thing, the albumins of the two mormoopid genera are, on the average, somewhat more distant from each other than either is from albumins of the phyllostomids. However, we would have expected, given the usual association of Mormoops and Pterono¬ tus on the basis of anatomical similarities, to find that their albumins had changed to a greater degree than those of the phyl¬ lostomids. Otherwise, the albumin data would not be readily interpretable within the generally agreed upon framework that 10 OCCASIONAL PAPERS MUSEUM TEXAS TECH UNIVERSITY immunological distances of Table 2. The numbers along the lines represent the amount of albumin change allocated to a particular lineage. The Family Pteropo- didae is used as an outside reference point to root the tree and apportion the amounts of change among lineages. represents mormoopids as a derived unit relative to other bats. Yet we find that the least-changed albumin in this set is that of Mor- moops me galop hylla, at a distance of 140 units from our ptero- podid reference set, whereas Pteronotus parnelhi is 178 and the phyllostomid mean is 172 (Table 2). These data may be explained most parsimoniously by suggesting that the phyllostomids and Pteronotus shared a common ancestor (separate from Morrnoops) in which there was an albumin change of approximately 30 or so AID units (Fig. 4)—in other words, a second event of similar magnitude as that proposed to have occurred in the common ancestral stock of all phyllostomoid bats. Noctilio is still another problem immunologically. Its albumin is divergent in the same way as that of other phyllostotnoids, being some 180 units from those of the pteropodids (Table 2). It also is quite different from those of other phyllostomoid lineages with which we are concerned. The distances separating Noctilio from Pteronotus, Morrnoops, and an assortment of phyllostomids are 150, 146, and 135 units, respectively (Table 2). The cladistic implications of these data are not unequivocal, For example, the apportionment of the Nochbo-phyllostomid distance of 135 units, using the pteropodid information, would suggest a Noctilio separa¬ tion somewhere along the common phyllostomoid lineage to which we have allocated a singular conformation-altering muta¬ tion (Fig. 4). Subsequent to that separation, it is then evident that this “event” was completed in a somewhat different fashion along ARNOLD ET AL.—PHYLLOSTOMOID BAT PHYLOGENY the Noctiho line relative to what happened along the Pteronotus- phyllostomid line, thus resulting in a markedly divergent albumin. Albumin Irregularities Over the last 15 years, extensive evidence has been developed that documents the generally time-dependent nature of immuno- logicallv assessed albumin evolution. However, significant indi¬ vidual departures from this pattern do exist in that a few lineages among many have accumulated change at a rate significantly greater or less than the average. Examples are Aotus, Caluromys, Marmosa, and Ursus on the slow side, and Rousettus and Phaner on the fast side (Sarich, 1969; Cronin and Sarich; Maxson et al., 1975). Of course, such a lineage ultimately could develop into a major adaptive radiation leading to a large clade, all the individ¬ ual lines of which might appear removed from clocklike behavior; one such instance already has been reported (Cronin and Sarich, 1975). In that particular case, it appears that at some time along the ancestral anthropoid lineage, subsequent to the anthropoid- prosimian split, about 25 to 30 units of change in excess of the average accumulated. Almost all anthropoid albumins thus appear to be changed to a greater degree than do those of most prosimians. Although this could be regarded simply as twice the usual rate of change over the period of time involved (about 30 million years), the rate being normal before and after, it is appre¬ ciably easier to envision it as resulting from a single mutation that somehow altered the conformation of the albumin surface so as to be the antigenic equivalent of many individual amino acid substitutions. There is excellent evidence that the number of differences in the surface amino acid sequence between two native proteins can be closely approximated by quantitative immunological comparisons (reviewed in Wilson et al., 1977), although alterations of the three- dimensional structure can have drastic effects. One can imagine, therefore, a single internal amino acid substitution (perhaps involving a cysteine and the subsequent relocation of one or more disulfide bridges) that could have a similar effect on the three- dimensional protein structure on a reduced scale. One also might interject a cautionary note here concerning the possible nonequi¬ valence of immunological distances derived from conformational changes and those derived from the accumulation of single amino acid substitutions. We include this as a final note for considera- 12 OCCASIONAL PAPERS MUSEUM TEXAS TECH UNIVERSITY tion of albumin evolution in phyllostomoid bats, where more than the usual number of interpretive problems exist. Character State Evolution and the Derivation of Phylogenies As witnessed by a continuing list of authors (Mickevich and Johnson, 1976; Schnell et al., 1978; Turner, 1974), the derivation of a consistent (that is, congruent) phylogeny using different character states is, to say the least, complicated. Is a similar method of analysis (cladistics) the key to deriving congruent phylogenies? Consistency seems achievable if one adequately can interpret patterns of character state divergence in terms of a sim¬ ilar method of analysis that potentially can detect differential rates of change, degree of homoplasy, and primitive (as opposed to derived) conditions. The cladistical method seems appropo; how¬ ever, regardless of the method of analysis, certain limitations inherent to a particular type of character may limit congruence (Fig. 5). In our data sets, limitations for different character states can be categorized as follows: 1) In terms of morphology, chirop- teran systematics is still at a stage where evolutionary relation¬ ships are based mainly on a “gestalt,” due in part to the lack of a fossil record. An appropriate quantitative approach (void of size relationships) to morphological relationships within chiropterans does not exist. The qualitative approach used in our study does support the phyllostomoid superfamilial association; however, the associations among the three families as well as those between Mormoops and Pteronotus are much more tentative (Figs. 2 and 5). 2) The electrophoretic approach is limited by inability to resolve synapomorphic character states among the genera Mor¬ moops, Pteronotus, and Noctilio (Fig. 1). 3) The chromosomal approach is limited because of our current inability to decipher chromosomal homologies between superfamilies and thus to determine conclusively (without the rule of parsimony) primitive conditions at superfamilial levels (Figs. 3 and 5). 4) Phylogenetic associations such as those implied with albumin immunological data are clearly more accurate in cases where divergence can be correlated to surface amino acid substitutions. Structural changes of the albumin molecule only can be inferred and at this stage in scientific investigations are not usually verifiable (Table 2). The inconsistent resolving power associated with a given suite of characters indicates that the best phylogeny generated from dif¬ ferent types of characters need not reflect total congruence. Rather, an alternate consideration would be one of compatibility. ARNOLD ET AL.—PHYLLOSTOMOID BAT PHYLOGENY 13 Fig. 5.—Composite cladogram indicating the levels of resolution of each of the character sets used: A=albumin, C^chromosomal, E —electrophoretic. and M= morphological. Every character does not have to resolve the same branching sequences at different evolutionary levels; however, all characters should lead to maximum compatibility. The failure of certain characters to resolve branching sequences at a given level then can be regarded as neutral insofar as determination of a consistent phylogeny. Incompatibility occurs only when different characters 14 OCCASIONAL PAPERS MUSEUM TEXAS TECH UNIVERSITY reveal conflicting branching sequences. One is then forced to assess this incompatibility in terms of the characters used and the inconsistencies and limitations associated with those characters (see immunological discussion). Our study supports the conclusion that a complex phylogeny can be resolved best by the use of multiple, differentially resolving character sets. Acknowledgments We thank Dr. A. C. Wilson for providing laboratory space for the immunology experiments. For assistance in collecting speci¬ mens, we thank H. H. Genoways, J. A. Groen, P. V. August, J. W. Bickham, and I. F. Greenbaum. O. F. Francke, M. W. Haiduk, L. W. Robbins, and R. L. Robbins reviewed various drafts of the manuscript. This work was supported by a grant from The John D. Archbold Family Trust, NSF grants DEB-76-20580 and DEB- 80-04293, and by a graduate student-faculty grant from the Grad¬ uate School, Texas Tech University . Literature Cited Baker, R. J., and J. W. Bickham. 1980. Karyotypic evolution in bats: evidence of extensive and conservative chromosomal evolution in closely related taxa. Syst. Zool., 29:239-253. Cronin, J. E., and V.. M. Sarich. 1975. Molecular systematics of the New World monkeys. J. Human Evol., 4:357-375. Dobson, G. E. 1875. Conspectus of the suborders, families, and genera of Chi- roptera arranged according to their natural affinities. Ann. Mag. Nat. Hist., ser. 4, 16:345-357. Hennig, W. 1966. Phvlogenic systematics. Univ. Illinois Press, Urbana, 274 pp. Maxson, L. R., and A. C. Wilson. 1974. Convergent morphological evolution detected by studying proteins of tree frogs in the Hyla eximia species group. Science, 185:66-68. Maxson, L. R., V. M. Sarich, and A. C. Wilson. 1975. Continental drift and the use of albumin as an evolutionary clock. Nature, 255:397-400. Mickevich, M. F., and M. S. Johnson. 1976. Congruence between morphologi¬ cal and allozyme data in evolutionary inference and character evolution. Syst. Zool., 25:260-270. Miller, G. S., Jr. 1907. The families and genera of bats. Bull. U.S. Nat. Mus,, 57:xvii+1-282. Patton, J. C., and R. J. Baker. 1978. Chromosomal homology and evolution of phyllostomatoid bats. Syst. Zool., 27:449-462. Prager, E. M., and A. C. Wilson. 1971. The dependence of immunological cross-reactivity upon sequence resemblance among lysozymes. J. Biol. Chem., 246:7010-7017. ARNOLD ET AL.—PHYLLOSTOMOID BAT PHYLOGENY 15 Prager, E. M., and A. C. Wilson. 1978. Construction of phylogenetic trees for proteins and nucleic acids: empirical evaluation of alternative matrix methods. J. Mol. Evol., 11:129-142. Sarich, V. M. 1969. Pinniped origins and the rate of evolution of carnivore albumins. Syst. Zool., 18:286-295. Sarich, V. M., and A. C. Wilson. 1966. Quantitative immunochemistry and the evolution of primate albumins: micro-complement fixation. Science, 154:1563-1566. Schnfll, G. D., T. L. Best, and M. L. Kennedy. 1978. Interspecific morpho¬ logic variation in kangaroo rats ( Dtpodomys): degree of concordance with genic variation. Syst. Zool., 27:34-48. Selander, R. K., M. H. Smith, S. Y. Yang, W. E. Johnson, and J. G. Gen¬ try. 1971. IV. Biochemical polymorphism and systematics in the genus Peromyscus. I. Variation in the Oldfield Mouse (Peromyscus polionolus). Studies in Genetics VI., llniv. Texas Publ., 7103: 49-90. Simfson, G. G. 1945. The principles of classification and a classification of mammals. Bull. Amer. Mus. Nat. Hist., 85:xvi+1-350. Smith, J. D. 1972. Systematics of the chiropteran family Morrnoopidae. Misc. Publ. Mus. Nat. Hist., Univ. Kansas, 56:1-132. Trocessart, E. L. 1897. Catalogue mammalium tarn viventum quam folilium. Nova edito {prirna completa), Berlin, R. Friedlander and Sohn, 1:1-218. Turner, B. J. 1974. Genetic divergence of Death Valley pupfish: biochemical versus morphological evidence. Evolution, 28:281-294. Wilson, A. C., S. S. Carlson, and T. J. White. 1977. Biochemical evolution. Ann. Rev. Biochern., 46:573-639. Winge, H. 1892. Jordfundne og nulevende Flagermus ( Chiroptera) fra Logoa Santa, Minas Geraes, Brasilien. Copenhagen, Museo Lundii, 2:1-65. Addresses of authors.— Arnold, Honeycutt, Baker, Jones: The Museum and Department of Biological Sciences, Texas Tech University, Lubbock, Texas 79409-, Sarich: Departments of Anthropology and Biochemistry, University of California, Berkeley 94720. Present address for Arnold and Honeycutt: Population Biology, Research School of Biological Sciences, The Australian National University, Can¬ berra City, A.C.T. 2601, Australia. (Direct reprint requests to The Museum.) Sub¬ mitted 1 May, accepted 30 July 1981. PUBLICATIONS of the museum TEXAS I'ECH UNIVERSITY Three publications of The Museum of Texas Tech Uni¬ versity are issued under the auspices of the Dean of the Graduate School and Director of Academic Publications, and in cooperation with the International Center for Arid and Semi-Arid Land Studies. Short research studies are published as Occasional Papers whereas longer contributions appear as Special Publications. Papers of practical application to col¬ lection management and museum operations are issued in the Museology series. All are numbered separately and pub¬ lished on an irregular basis. The preferred abbreviation for citing The Museum's Occa¬ sional Papers is Occas. Papers Mus., Texas Tech Univ. Institutional subscriptions are available through Texas Tech Press, Texas Tech University, Lubbock, Texas 79409. Institutional libraries interested in exchanging publications should address the Exchange Librarian at Texas Tech Uni¬ versity. Individuals can purchase separate numbers of the Occasional Papers for $1.00 each from Texas Tech Press. Remittance in U.S. currency check, money order, or bank draft must be enclosed with request (add $1.00 per title or 200 pages of publications requested for foreign postage; resi¬ dents of the state of Texas must pay a 5 per cent sales tax on the total purchase price). Copies of the “Revised checklist of North American mammals north of Mexico, 1979” (Jones et al ., 1979, Occas. Papers Mus., Texas Tech Univ., 62:1-17) are available at 60 cents each in orders of 10 or more. ISSN 0149-175X Texas Tech Press Lubbock, Texas 79409